Concealed weapons pose a significant threat to both law enforcement and security agency personnel. The uncontrolled environments associated with peacekeeping and the move toward relaxation of concealed weapons laws here in the U.S. provide a strong motivation for developing weapons detection technologies which are noninvasive and can function noncooperatively. Existing weapons detection systems are primarily oriented to detecting metal and require the cooperation of the person being searched. The new generation of detectors under development that focuses primarily on imaging methods, faces problems associated with privacy issues. There remains a need for a weapons detector which is portable, detects weapons remotely, avoids the issues associated with privacy rights, can tell the difference between car keys and a knife, and is affordable enough that one can be issued to every peacekeeper and law enforcement officer. AKELA is developing a concealed weapons detector that uses wideband radar techniques to excite natural electromagnetic resonances that characterize the size, shape, and material composition of an object. Neural network processing is used to classify the difference between weapons and nuisance objects. We have constructed both time and frequency domain test systems and used them to gather experimental data on a variety of armed and unarmed individuals. These experiments have been performed in an environment similar to the operational environment. Preliminary results from these experiments show that it is possible to detect a weapon being carried by an individual from a distance of 10 to 15 feet, and to detect a weapon being concealed behind the back. The power required is about 100 milliwatts. A breadboard system is being fabricated and will be used by AKELA and our law enforcement partner to gather data in operationally realistic situations. While a laptop computer will control the breadboard system, the wideband radar electronics will fit in a box the size of a CD ROM drive of a computer.